Chair: D. B. Haidvogel, Institute of Marine and Coastal Sciences, Rutgers University, New Brunswick, NJ 08901

Participants: Mark Abbott, Jim Bellingham, Yi Chao, Ken Denman, Bruce Frost, Barbara Hickey, Zack Powell, Steve Riser, Rick Thomson

1. Background

Over the past few decades, our knowledge of ocean circulation and variability has advanced significantly thanks to the several international ocean observing programs, e.g., TOGA, WOCE, JGOFS, and GLOBEC. Nonetheless, until recently, in situ ocean observing systems have been limited in their spatial and temporal coverage, e.g., to single-point moorings (like BATS and HOT), to regional networked moorings (TAO and PIRATA), and to sparsely distributed surface drifters and subsurface floats. Although satellite remote-sensing technology has progressed to a point at which almost all surface variables (except salinity) can be routinely measured from space, the oceanic interior has been largely unsampled, except in a few small (predominantly coastal) regions.

Recent technological advances in sensors and observing systems on the one hand, and numerical models and data assimilation on the other, have begun to change this situation dramatically. Observing networks combining remote sensing datasets with extensive subsurface fields obtained from both fixed and autonomous platforms are being exploited in several regions. Modern communication systems and the World Wide Web enable real-time reporting and distribution of these data. And finally, numerical ocean circulation models, and associated techniques of data assimilation, now allow meaningful forecasts of future states of the physical circulation, and have begun to provide hindcasts and forecasts of coupled chemical and biological processes. Recent reviews describing progress in these areas can be found in Glenn et al. (1999) and Haidvogel et al. (1999).

The NEPTUNE program offers exciting opportunities to move ahead in three broad areas. First, the NEPTUNE system as envisioned will enable the deployment of an observing network of unprecedented density, variety and spatial coverage, including novel instrumentation for sensing the combined in situ physical-chemical-biological environment. Second, assimilative models can be applied to design the NEPTUNE physical-chemical-biological observation system (taking advantage of the ability to sample completely a "block" of ocean in four dimensions), to produce forecasts of episodic oceanic events, and to adaptively redirect the deployment of moveable observational platforms in response. Third, the availability of a unified observing/modeling system for the NEPTUNE region will allow scientists to address issues of physical/biological coupling across multiple spatial and temporal scales.

2. Theme 1: Observational technologies

Since uniform sampling at high spatial and temporal resolution over the entire NEPTUNE domain is impractical, and given that the critical processes often occur only sporadically, a dynamical and adaptive sampling system will be required. In the long term, the resulting dynamic sampling strategy would merge both modeling and observing systems to provide a predictive capability for both physical and biological variables. Since predictability timescales are short for small-scale physical variables of interest (e.g., fronts and eddies), and presumably even shorter for chemical and biological variables, it will be necessary to constrain numerical models using real-time datasets.

Table 1. Status of various satellite remote sensing datasets

Variable	satellite program	status
ocean color temperature sea level bottom pressure	SeaWiFs AVHRR TOPEX/POSEIDON GRACE	since 1997 since 1980s since 1992 2001

As they have in the past, datasets obtained remotely from space-, land-based and in situ platforms will play an essential role in the NEPTUNE observing system. In addition to records of sea surface temperature, ocean topography, ocean color, and vector winds, new satellite measurements are also being proposed (Table 1). In particular, sea surface salinity may be measured from satellites in the next decade. Complementing these satellite systems (which "see" only the ocean surface) will be other platforms capable of providing spatial patterns of oceanic fields, including land-based, low-frequency HF radar platforms (surface velocity); acoustic arrays (ocean temperature; e.g., Fig. 1); and freely floating and autonomous floats, gliders and AUV's carrying a wide range of sensors (velocity, T/S, turbulence, etc.).

Contemporaneous in situ measurements are necessary not only for the independent information they give on (e.g.) the ocean interior, but will also be required as independent checks (ground-truth) for the satellite measurements. Ground truthing of this and other new observing techniques is necessary to demonstrate their utility for ocean research. Spatially extensive, high-resolution ground truthing for these systems could be a major contribution of the NEPTUNE program, making the water column measurements made by NEPTUNE as important to the satellite missions as the satellite data is to NEPTUNE.

To realize the full potential of NEPTUNE will require the development of "self sufficient" samplers: all remote optical sensors will require self-cleaning capability - feasible with NEPTUNE because of assured power supply. Adaptive purposeful sampling triggered by real-time observing will allow comprehensive calibration and validation of continuous sensors not presently possible. The ocean is also incredibly dilute and NEPTUNE will allow concentration (of plants, animals, and chemicals) through filtering, now only possible with great effort via strings of pumps, and through flow cytometry, now possible only from ships, except with great effort.

Exploitation of NEPTUNE will require that instrumentation move into "commodity" space where costs are low, and the value of the network is driven by the services which are provided. Present generation ocean sensors are generally special purpose and production numbers are small. In such an environment, the industry is dominated by small companies which often find it difficult to raise the necessary capital to support research and development. Unlike the PC industry, the sensor industry has not been able to drastically reduce profit margins and expand production. As the Internet has shown, the value of the network increases exponentially with the number of connected devices. NEPTUNE must overcome these barriers of cost as well as capability. The nodes on the network must become the elements of intelligence, not simply data collection devices relaying their information to central servers.

The investment required is substantial; it goes beyond incremental improvements to the existing set of sensors. The need for laboratory development, extensive pre-deployment testing, and a tolerance for failure is essential. However, much of the necessary technology will come from outside the sensor industry. Thus it is essential that some of the development support go towards tracking and implementing new capabilities that come from other sectors. The point is that other technical sectors will lead the effort, and the ocean sensor industry must be prepared to infuse these developments.

3. Theme 2: NEPTUNE Ocean Prediction System (NOPS)

Given the proposed NEPTUNE observing system, a real-time, or near-real-time, ocean prediction system (NOPS) will be feasible. Elements of the system would likely include regional oceanic/atmospheric model, initialized with NEPTUNE datasets, and forced on its lateral and top boundaries by larger-scale (basin-to-global) atmospheric and oceanic circulation models. Data assimilation software would be used to combine observations from NEPTUNE platforms with model-based estimates in a dynamically appropriate way. Eventually, adaptive refinement of the observing system ( e.g., the AUV's) could be implemented. The resulting observational/modeling system would be used to predict the physical character of the water column, and to provide the context for observing and understanding water column processes across a wide range of space/time scales encompassing episodic, seasonal and decadal variability. A schematic of such a system is shown in Fig. 2.

A variety of episodic processes are either partially or completely within the water column. An important, and poorly observed, example of these are seafloor eruptions. While these events can be detected acoustically or seismically, their water column signature consisting of heat and particulate matter (both organic and inorganic) might be the first direct verification that such an event has occurred. One important mission for NEPTUNE would therefore be to detect and to characterize such plumes, with the NOPS providing guidance on plume location, spreading rate, etc.. Important questions include: how much heat was released, what are the microbial signatures associated with the event, what chemicals were released, does the plume segregate itself, how high in the water column does it penetrate, and what is the time evolution as the event progresses? "Smart" autonomous underwater vehicles (AUVs), and new sensors for (e.g.) biological populations, would play an important role in such missions.

Ocean currents, particularly those on or near the continental shelf, play a prominent role in controlling coastal climate and shelf ecosystems (Brink and Robinson, 1998). The California Current that drives most of the flow along that coast originates in the Northeast Pacific, where the NEPTUNE array will be located. In the NE Pacific, the North Pacific Drift current (fed by the Kuroshiro off Japan) splits, with the bulk of the water going southeast in the California Current and the rest going northeast into the Alaskan Gyre. The exact amount each receives may be related to the timing of the long-term cycles in coastal fisheries in both Alaska and California. However, there are no regular current observations available in the NE Pacific. NOPS would provide the capability of monitoring, modeling, and dynamically interpreting the actual transport of each portion of the split. Taken together with other long-term and/or ongoing data collection efforts - e.g., CalCOFI, GLOBEC, etc. - we are in a position to achieve an integrated picture of the whole California Current system.

Recent studies have shown that there is a basin-wide interdecadal climate oscillation in the Pacific Ocean (Mantua et al., 1997). For example, the well-known 1976-77 climate regime shift probably represents the phase transition of this nearly periodic interdecadal oscillation. This Pacific Decadal Oscillation (PDO) is characterized by cooling in the western and central North Pacific and warming in the eastern equatorial Pacific and also along the coasts of North and South America. Thus, the proposed NEPTUNE system is sited within a key region of this see-saw, basin-wide pattern. The long-term deployment of NOPS in this region would provide an unprecedented description of the regional expressions of these basin-wide climate phenomenon on multi-decadal time scales. The signature of the PDO in the equatorial Pacific Ocean is also believed to have a major impact on the decadal variability of ENSO (e.g., McPhaden, 1999), a central theme to be explored by the CLIVAR program.

4. Theme 3: Physical-Chemical-Biological "connections"

Modeling and prediction of ocean chemistry and biology would also be an important goal of the NOPS system. (See. e.g., Fig. 3 for a schematic of one existing coupled physical/biological modeling system.) One significant limitation on our current capability to model ecosystems is the relative lack of tested theories and quantitative measures of biological rates and processes. The NEPTUNE system offers an important prospect for making significant advances in this area.

The ocean is an extremely rich habitat for micro-organisms, and there appears to be much higher species variability both spatially and temporally than observed for the physical parameters of seawater. This variability is poorly understood. Fortunately, there are a number of scientists developing in situ sensors for both the chemical nutrients and for the micro-organisms themselves. As these systems become available, there will be a wide range of difficult scientific challenges associated with developing and understanding the interdependencies of ocean physical parameters, chemistry, biology, and seafloor/atmosphere interactions. The characteristic time scales of phytoplankton growth (doubling times of one to several days) result in a close coupling with mesoscale variability of physical processes through its impact on light and nutrient availability (e.g., McGillicuddy et al., 1999). Our present understanding of this coupling is largely through episodic or opportunistic sampling; long-term, consistent time series of both the biological and physical processes at the appropriate time and space scales is largely absent.

There is accumulating evidence that the structure of planktonic populations is at least as important in determining the future behavior of those populations as are the more difficult to measure rates of growth, death, grazing, etc. Zooplankton structure can be monitored by both Video Plankton Recorders and by multifrequency acoustical sensors. Both have high power requirements and high data storage requirements - and are usually used separately and for relatively short periods. NEPTUNE (because of power availability and data telemetering) offers the possibility of dense sampling (in time and space) of the zooplankton community both optically and acoustically at the same time. Real-time observing also allows for initiation of purposeful sampling with remote nets or traps. With SeaWiFS and subsequent multichannel satellite color sensors like MODIS, we have the potential ability to probe the structure of phytoplankton populations, but developing the ability to determine remotely the community structure of phytoplankton will require extensive in situ monitoring and calibrating. NEPTUNE offers an unparalleled opportunity to obtain high quality simultaneous in situ color spectral information and species composition information and samples. This will both greatly accelerate the development of the capability to determine phytoplankton community structure from satellite and also provide otherwise unavailable high quality in situ data on community structure for marine ecosystem models, both for understanding and for prediction.

The fluxes of heat and chemical and biological species (e.g., CO2) into and out of the ocean are poorly understood. One approach to these fluxes is to create an observational box in the ocean wherein one can carefully monitor all boundary exchanges. An underwater laboratory such as NEPTUNE -- along with an accompanying physical/biological modeling system with hindcast and (eventually) predictive capabilities - provides the ideal logistical base for supporting such studies. For example, there is great controversy relating the production of sinking organic particles in the surface euphotic zone and the capture of organic particles at depths of 150 to 300m below the euphotic zone. One reason is the "catchability" of traps and the second reason, especially in continental margins, is lateral transport of sinking particles. A NEPTUNE network/grid of water column current meters would map the current field continuously at spatial resolutions heretofore impossible. Sinking particles could be captured and/or tracked optically or chemically, and the Lagrangian motions and mixing of the water column could be simultaneously observed with purposeful release of dye or chemical (e.g., SF6). To a large degree, the big problem of JGOFS -- the sequestering of organic carbon particles at depth -- remains unresolved, and NEPTUNE could go a long way towards resolving it.

References

Brink, K. and A. R. Robinson (1998). The Sea, 10.

Mantua et al., Bull. Am. Meteorol. Soc. 78, 1069 (1997).

McGillicuddy et al. (1999).

McPhaden, Science, 283, 950, 1999.